Kevin Ashley, Head of Digital Archives, NDAD and Jeffrey Darlington, Digital Preservation Systems Manager, Public Records Office

The National Digital Archive of Datasets (NDAD) provides a service to the Public Record Office of England and Wales to acquire, preserve, catalogue and provide access to structured digital records from UK government - what most people would call 'databases'. Amongst the accessions we've dealt with recently is SPRINT (Single Professional Register and Index of Training) from the UK Central Council for Nursing and Midwifery. Regulation of nursing and related professions has recently undergone major revision in the UK, with three separate statutory bodies being wound up on 31 March 2002 and replaced by a new single body (the NMC) on 1 April 2002. The UKCC was one of the bodies wound up and as is often the case, review of records whose life would be at an end following the merger was left to a late stage, even more so with the computer-based records. We only had a few days in between hearing about this database and arranging its capture for preservation.

SPRINT is an aging ADABAS database which has been replaced following the creation of the new regulatory body by a more modern system which covers a much wider set of functions, but not all of those which were covered by SPRINT. The changing regulations mean that some of the information SPRINT holds is no longer required. If the information unique to SPRINT was deemed worthy of preservation it thus had to be captured quickly. Since all IT provision was outsourced, there was little or no ability to capture data from the system once the outsourcing contract terminated.

Outsourcing is not uncommon with government IT and presents additional problems to those trying to preserve records that come under its remit. In this case, it meant an extra organisational layer between NDAD (trying to preserve the records and their metadata) and the records themselves. One of the results of this was that, despite the fact that ADABAS contained machine-readable metadata such as field names and data types which are essential to us, our contact at the outsourcing company had felt it necessary to transcribe all that information, by hand, into a Microsoft Word document which was all that they would supply us with in the way of metadata. The cost, to them and us, of this approach is considerably more than if we were able to access the records directly ourselves.

Nonetheless we're now confident we have enough to begin the task of preservation and description. The catalogues and dataset, which continues a series of paper professional registers stretching back many years, will become available on the NDAD website (ndad.ulcc.ac.uk).

As with all of NDAD's accessions, the finding aids and the records themselves are closely linked allowing one to move back and forth between the records (be they documents, database tables or images), their descriptions, and associated records. This is achieved via an integrated metadata management system which automatically generates finding aids for the most detailed units of description (with embedded links to the resources they describe) and also manages access to the records themselves.

Every silver cloud has a dark lining, however. Although the metadata database allows automated management of many aspects of the archive, from scheduling consistency checks through production of finding aids and automatic release of closed material, the capture of that metadata is itself anything but automatic. It is still a matter of some mild astonishment to us that, in many cases, the only metadata for a large and complex record collection exists as a poor fifth-generation paper copy of a printout, or worse only as a set of handwritten notes. The focus of past custodians was often on data preservation rather than record preservation. The result has sometimes been that we have all the numbers, but little indication of what those numbers are trying to tell us or why. For the bulk of NDAD's holdings, this is not the case; tracking down that metadata has not always been easy but we've managed - so far.   º

JEFFREY DARLINGTON, PUBLIC RECORD OFFICE, ADDS: These problems occur in spite of the advice that the Public Record Office provides to government departments on managing their information assets. Records management standards for electronic records often fall far short of those achieved for paper records, and IT outsourcing sometimes creates a barrier that prevents the advice from getting to where it is most needed. We are developing further guidance on how to ensure that the records management requirement is addressed in the design of new databases, and also how existing databases can be better managed. We hope that our educational activities will raise standards and lead to trouble-free accessions of electronic records in future.  º